Alzheimer's disease (AD) is preceded by a period when detectable changes in brain function occur without warning symptoms. This period may be twenty years or longer. Activation measured by functional magnetic resonance imaging is reduced during confrontation naming in normal women who have increased risk of late-onset Alzheimer's disease, at an average age of only 53 years. These same individuals have reduced activation in the posterior cingulate cortex during a memorization task, among other disparities. Our hypothesis is that the altered functional MRI responses in naming, working memory and memorization will quantifiably worsen with age, due to progressive underlying Alzheimer's disease pathology. Disease-modifying treatments applied in this early, pre-symptomatic stage of AD could have profound impact, by preventing the onset of cognitive symptoms. Millions are currently being spent on large-scale prevention trials, with AD symptoms as end-points. By providing a biomarker of pre-symptomatic AD progression, fMRI could potentially reduce the duration and costs of such trials. This continuation proposal is designed to detect changes in brain function in high-AD risk individuals over time. We will study normal education-matched groups of high- and low-AD risk subjects in the age ranges 40-65 and 65-90 years, using fMRI stimulus tasks which have previously demonstrated regional disparities in high-AD risk individuals. In addition, we will repeat the naming and fluency fMRI studies performed previously in high- and low-AD risk individuals after an interval of five years, in order to detect longitudinal changes in activation. The convergence of evidence from these cross-sectional and longitudinal studies could provide powerful evidence for a model of Alzheimer's disease as a relentless, slowly progressive brain pathology that begins early in adult life, but remains compensated until it produces clinical symptoms in its late stages.